April 09, 2009

Serendipitous Double Entendre

Search engines are generally about the words we use. If you use the words "second base", "foul", and "home", you're probably writing about (or searching for) baseball. The current state of the web (and of search) doesn't have any context; it only has the sequence of letters in a search string and maybe a lexicon of similar words for substitution, but at this time search (and the Web) doesn't have any grasp of the meaning of the words. The meaning of "fanny", for instance, from the picture at left, is a shared concept within all of our heads, but the Web just sees it as a five-character alphanumeric string. (which goes back to Shannon and Weaver's theory of information, see this earlier post).

The Semantic Web (youtube), which some people call Web 3.0, hopes to be all about meaning. In today's system, based on words as text strings without meaning, there's a possibility of confusion. The sporting side of me knows that sometimes confusion is an opportunity.

I spend a bit of time studying and learning about Search Engines (primarily Google) because I have SEO (search-engine-optimization) clients. I run little SEO experiments. Nothing a mad scientist would get excited about; no lightning rods and thunderstorms. There's a significant time latency and it's a multi-variable problem; it's difficult to measure the effect of a single change because while I'm changing the website, Google is tweaking their code, too.

Yesterday I stumbled across a serendipitous double entendre, that is to say an unexpected double meaning that produced excellent search results in an unanticipated way. I found it in my blog's SiteMeter records: Google brought me somebody from London who was searching for tawdry naughtiness.

SiteMeter is a blog service which allows you to see some high-level details (like general location) about your visitors. For instance, my last ten visitors have been from:
       Madras, Tamil Nadu, India
Cecil, Pennsylvania
Huntersville, North Carolina
Cecil, Pennsylvania
Pittsburgh, Pennsylvania
 Tucson, Arizona
London, England
New City, New York
Oosterzele, Oost-Vlaanderen, Belgium
Norwalk, Connecticut

You can really only tell where their internet connection or cable company comes from; there's no street addresses. Data from server logs is interesting in an meta-data aggregate way, but pretty much useless in an individual sense. You can also tell how people came to be on your site. This is called referral data - what other site referred them to your site?

I had a visitor yesterday from England (west side of London) and the referral info indicated he had Googled (via Google.co.uk, the United Kingdom variation) "flirty temptations" and he ended up at my site. So I googled the same term, in both the UK and the Stateside Google, and remarkably this blog is #1 in Google for "flirty temptations", in spite of several businesses that use that phrase on their websites.

This is very amusing. Last month I wrote about how we'd had a stretch of mild weather and my first bike ride of the season, and I titled it "The Flirty Temptations of Imminent Spring". This was not an experiment, it was just a blog entry about a nice day. Sometimes a cigar is just a cigar.

I can only imagine an Englishman on the west side of London googling "flirty temptations". I have to imagine he was disappointed to find a story about the Montour Trail.

0 comments:

Post a Comment

Comments and Feedback? Love that stuff. Please leave your thoughts in the box below--